Non-Gaussian Chi-squared method with the multivariate Edgeworth expansion

摘要: I present here a generalization of the maximum likelihood method and $\chi^2$ to cases in which data are {\it not} assumed be Gaussian distributed. The method, based on multivariate Edgeworth expansion, can find several astrophysical applications. mention only two them. First, microwave background analysis, where it cannot excluded that initial perturbations non-Gaussian. Second, large scale structure statistics, as we already know galaxy distribution deviates from Gaussianity scales at non-linearity is important. As first interesting result show how confidence regions modified when non-Gaussianity taken into account.